mcp-server-docx
Enables creating professional Word documents from markdown or structured content with fast, customized formatting via natural language.
README
Word Document MCP Server
A fast, TypeScript-based MCP server for creating professional Word documents from markdown or structured content.
š Markdown-First Workflow - Just write natural markdown and get professional Word documents in ~300ms!
Quick Setup
-
Download
index.jsfrom Releases and save to~/mcp-servers/docx/index.js -
Configure Claude Desktop - edit
claude_desktop_config.json:- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
{ "mcpServers": { "mcp-server-docx": { "command": "node", "args": ["/Users/YOUR_USERNAME/mcp-servers/docx/index.js"] } } }Replace the path with where you saved the file (Windows: use
C:\\Users\\...with double backslashes) - macOS:
-
Restart Claude Desktop
-
Test by asking Claude:
Create a Word document at /tmp/test.docx from this markdown:
# Hello World
This is my **first** Word document with *markdown*!
## Features
- Easy to use
- Fast generation
- Professional formatting
You should get a properly formatted Word document at /tmp/test.docx!
<details> <summary><b>Build from Source</b> (for contributors)</summary>
git clone <repository-url>
cd mcp-server-docx
nvm install && npm install && npm run build
Use dist/index.js in your Claude config instead of the downloaded bundle.
</details>
Features
šÆ Pure Markdown Mode (Recommended)
The simplest way to create Word documents - just write natural markdown like you always do!
Just tell Claude what you want in plain markdown syntax:
Create a Word document at /tmp/my-resume.docx from this markdown:
# JANE SMITH
jane@example.com | (555) 123-4567
## PROFESSIONAL SUMMARY
I am a **senior software engineer** with *10+ years* of experience building scalable web applications.
> "Jane is an exceptional technical leader and mentor." ā Former Manager
## SKILLS
- Expert in **TypeScript**, **React**, and **Node.js**
- Strong experience with **AWS** cloud architecture
- Passionate about **clean code** and **best practices**
---
## WORK EXPERIENCE
### Senior Engineer at Tech Corp (2020-Present)
Key achievements:
1. Designed and implemented **microservices architecture**
2. Reduced system latency by *40%*
3. Mentored team of 5 junior engineers
### Software Engineer at Startup Inc (2015-2020)
- Built MVP from scratch using **React** and **TypeScript**
- Implemented REST API with **Express** and **PostgreSQL**
What you get:
- ā Professional Word document in ~300ms
- ā Proper heading styles with borders (H1, H2)
- ā Bold and italic formatting
- ā Bulleted and numbered lists
- ā Block quotes formatted in italics
- ā Horizontal rules handled automatically
- ā Proper spacing between sections
- ā Times New Roman font (professional default)
Supported Markdown:
######Headings (H1/H2 get bottom borders)**bold**and*italic*inline formatting[text](url)for clickable hyperlinks-or*for bullet lists1.2.for numbered lists> quotefor block quotes (rendered in italics)---***___horizontal rules (rendered as lines)- Empty lines for spacing between sections
Behind the scenes: Uses create_document_from_markdown tool
šØ Custom Styling for Markdown
Want Helvetica instead of Times New Roman? Different font sizes? - Easily customize the appearance!
You can override the default styles by passing a styles object:
Create a Word document at /tmp/my-resume.docx from this markdown:
# JANE SMITH
## Professional Summary
I am a senior engineer.
Use these custom styles:
- All headings should be Helvetica
- H1 should be 36pt bold without bottom border
- H2 should be 24pt bold with bottom border
- Paragraphs should be Arial 12pt
Style options by element:
heading1,heading2,heading3,heading4- Control heading appearanceparagraph- Regular paragraph textbullets- Bulleted list itemsordered- Numbered list itemsblockquote- Block quote text (from> quote)
Each element can have:
fontName- Font family (e.g., "Helvetica", "Arial", "Times New Roman")fontSize- Font size in points (e.g., 12, 14, 36)bold- Bold text (true/false)italic- Italic text (true/false)color- Text color as hex RGB (e.g., "FF0000")borderBottom- Bottom border for headings (true/false)
Example styles object:
styles: {
heading1: { fontName: 'Helvetica', fontSize: 36, bold: true, borderBottom: false },
heading2: { fontName: 'Helvetica', fontSize: 24, bold: true, borderBottom: true },
heading3: { fontName: 'Helvetica', fontSize: 18, bold: true },
paragraph: { fontName: 'Arial', fontSize: 12 },
bullets: { fontName: 'Arial', fontSize: 12 },
ordered: { fontName: 'Arial', fontSize: 12 },
blockquote: { fontName: 'Georgia', fontSize: 12, italic: true }
}
Default styles (used when you don't provide custom styles):
- Headings: Times New Roman, bold, H1=24pt, H2=18pt, H3/H4=14/12pt
- H1/H2: Bottom borders, H3/H4: No borders
- Paragraphs/Lists: Times New Roman, 12pt
- Blockquotes: Times New Roman, 12pt, italic
Pro tip: You only need to specify the elements/properties you want to change! Unspecified elements use the defaults.
ā” Batch Mode with Structured Content
For when you need fine-grained control - specify exact fonts, sizes, and colors
Use this when markdown isn't flexible enough and you need precise control over formatting:
Create a resume at /tmp/my-resume.docx using create_document_from_content.
Make the name "JANE SMITH" in Helvetica 36pt bold.
Add a "PROFESSIONAL SUMMARY" H2 heading with a bottom border in Helvetica 14pt.
Add a paragraph about my experience in Times New Roman 12pt.
Add a "SKILLS" H2 heading with a bottom border.
Add bullet points for my skills in Times New Roman 12pt.
Example structure:
content: [
{
text: 'JANE SMITH', // defaults to paragraph type
format: { fontName: 'Helvetica', fontSize: 36, bold: true },
},
{ text: '' }, // empty paragraph for spacing
{
type: 'heading',
text: 'PROFESSIONAL SUMMARY',
format: { level: 2, borderBottom: true },
},
{
text: 'Software engineer with 10+ years experience...',
format: { fontName: 'Times New Roman', fontSize: 12 },
},
{
type: 'bullets',
items: ['TypeScript & React', 'Node.js & Python', 'AWS & Docker'],
format: { fontName: 'Times New Roman', fontSize: 12 },
},
];
Content item options:
type:'paragraph'(default),'heading','bullets','ordered'text: For paragraphs and headings (use''for spacing)items: Array of strings for listsformat:fontName,fontSize,bold,italic,color,level(headings),borderBottom(headings)
Behind the scenes: Uses create_document_from_content tool
š Using Markdown Formatting in Structured Content
You can use markdown-style formatting even in structured mode!
Both the pure markdown approach and structured content support inline **bold** and *italic* formatting:
// Markdown formatting works in any text field:
{
text: 'Led **4-engineer team** building *next-gen platform*',
format: { fontSize: 12 }
}
// Also works in list items:
{
type: 'bullets',
items: [
'Expert in **TypeScript** and **React**',
'Passionate about *clean code* and *best practices*'
]
}
This gets automatically converted to proper Word formatting with bold and italic runs.
Contributing & Local Development
Installation
git clone <repository-url>
cd mcp-server-docx
nvm install # Use the correct Node.js version from .nvmrc
npm install
npm run build
Creating a Release
This project uses automated GitHub releases. See RELEASING.md for details.
Quick summary:
- Update version in
package.json(e.g.,npm version patch) - Create PR and merge to
main - GitHub Actions automatically creates a release with the bundled
index.js
The release workflow only triggers on changes to:
src/**(excluding tests)package.jsonpackage-lock.json
Testing
Comprehensive test suite with 54 tests powered by Vitest:
npm test # Run tests once
npm run test:watch # Watch mode with fast HMR
npm run test:coverage # Coverage report
npm run test:ui # Visual UI mode
Test coverage:
- Markdown parsing (headings, lists, block quotes, horizontal rules, inline formatting, links)
- Auto-session creation
- Paragraph, heading, and list formatting
- Link support in paragraphs, headings, and lists
- Horizontal rule rendering
- Color formatting for all element types
- Batch document creation
- Error handling
- Complex multi-section documents
- Spacing between elements
Performance: All 54 tests complete in ~400ms
Code Quality
npm run lint # Run ESLint (fails on warnings)
npm run lint:fix # Auto-fix linting issues
npm run format # Format code with Prettier
npm run format:check # Check formatting (CI)
npm run typecheck # TypeScript type checking
npm run ci # Run all checks (lint + format + test + typecheck)
Standards:
- TypeScript with strict mode
- ESLint (no warnings allowed)
- Prettier (single quotes, 2-space indent, 100 char width)
- GitHub Actions CI on all PRs
Project Structure
mcp-server-docx/
āāā src/
ā āāā __tests__/
ā ā āāā document-manager.test.ts # Document management tests
ā ā āāā markdown-parser.test.ts # Markdown parsing tests
ā āāā index.ts # MCP server implementation
ā āāā document-manager.ts # Core document logic
ā āāā markdown-parser.ts # Markdown to content converter
ā āāā types.ts # TypeScript type definitions
āāā dist/ # Compiled JavaScript
āāā vitest.config.ts # Test configuration
āāā package.json
āāā tsconfig.json
āāā README.md
Performance & Architecture
Key Performance Metrics:
- Single document creation: ~300ms
- Full resume (batch mode): ~300ms vs 3-5s (incremental) vs ~35s (Python)
- 100x faster than Python implementation
- 20-30x fewer MCP calls (1 batch call vs 20-30 incremental)
Design Principles:
- Batch Operations - Create entire documents in a single MCP call
- In-Memory Storage - Documents stored in memory until save (eliminates file I/O overhead)
- Auto-Session Creation - No explicit initialization needed, start adding content immediately
- Smart Markdown Parsing - Single-pass parsing with inline
**bold**/*italic*support - Default Styling - Times New Roman applied automatically for professional appearance
- Type Safety - Full TypeScript definitions for correctness
Future Enhancements
- Tables and images
- Custom style definitions
- Headers, footers, and page breaks
- Search & replace functionality
License
ISC
Author
James Mehorter
Acknowledgments
This project is built on two excellent open source libraries:
Document Generation: docx
docx.js.org | GitHub - A powerful library for generating Word documents (.docx files) in JavaScript/TypeScript. This MCP server wouldn't be possible without the solid foundation that docx provides for creating and manipulating Office Open XML documents.
Markdown Parsing: remark
remark.js.org | GitHub - A robust markdown processor powered by plugins, part of the unified collective. Remark's standards-compliant parsing ensures accurate conversion of markdown to Word documents.
Thanks to:
- Dolan Miu and all contributors to the docx library
- Titus Wormer and the unified collective for remark
- The Anthropic team for building the Model Context Protocol and Claude Code
- The open source community for making tools like this possible
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.